OSF HealthCare

OSF HealthCare, a large integrated health system, manages thousands of purchased services contracts — many running hundreds of pages — covering vendors across facilities, equipment, and clinical support. Purchased services represent nearly half of non-labor spend for health systems, yet the vast majority of those contracts remain effectively unmanaged at the point of payment. Without automated enforcement, AP teams have no practical way to validate every invoice line against negotiated contract terms before funds leave the organization. The result is a systemic, industry-wide problem: more than $32 billion in preventable overpayments annually, driven by contracts that exist in isolation from the payment workflows they are supposed to govern.

OSF HealthCare deployed SpendRule, an AI-powered contract intelligence platform that uses large language models and generative AI to parse and encode contract terms, conditions, and pricing obligations into machine-readable payment controls. Rather than replacing existing infrastructure, SpendRule integrates directly into OSF's accounts payable workflows, sitting within the ERP environment already in use. At the moment an invoice arrives, the system performs 4-way matching — reconciling the Purchase Order, receipt record, invoice line items, and contract terms simultaneously. Discrepancies are flagged automatically with supporting evidence, routed for upstream resolution, and stopped before payment is released. This architecture converts what were previously static, siloed contract documents into active, real-time enforcement mechanisms.

Invoice validation at OSF HealthCare shifted from reactive auditing to proactive prevention. Key outcomes include:

• Thousands of manual approval hours eliminated per year, freeing supply chain staff to focus on exception handling rather than routine validation
• Invoice accuracy is now systematically enforced at the point of payment, eliminating the reliance on post-payment audits to recover overpaid amounts
• OSF gained visibility and control over a segment of spend — purchased services — that had previously been structurally difficult to manage at scale

The deployment demonstrates that AI-driven contract enforcement can close the gap between contracted terms and actual payments without requiring significant changes to existing ERP systems.

• Integration beats replacement: embedding contract enforcement inside existing AP workflows avoids the adoption friction of standalone contract management tools and delivers compliance at the moment payments are initiated.
• 4-way matching is the architectural standard: adding contract terms as a fourth match dimension alongside PO, receipt, and invoice is the key pattern for catching purchased-services overpayments before they occur.
• LLMs unlock previously unstructured contracts: AI's ability to parse dense, multi-hundred-page contracts at scale makes enforcement feasible where manual review is not.
• Purchased services demand dedicated tooling: generic spend management platforms are not designed for the complexity of services contracts — purpose-built solutions are necessary for meaningful coverage.